Fast computations of Wasserstein gradient flows
Abstract: Wasserstein gradient flows describe the evolution of a mass density flowing along a pressure gradient generated by an internal energy functional. This class of PDEs models various physical phenomena such as fluid flow, heat transfer, aggregation-diffusion and crowd motion. In general, these equations are both stiff and non-linear making them challenging to solve numerically. I will present a new algorithm for optimal transport allowing to efficiently run large scale gradient flows simulations for a large class of internal energies including singular and non-convex energies.
Joint work with Matt Jacobs and Wonjun Lee.
